A Hybrid Short-Term Load Forecasting Model Based on CatBoost and LSTM

In this paper, a hybrid model is proposed to forecast short-term load. The hybrid model consists of two layers. In the first layer, multiple CatBoost and LSTM models are trained using K-Fold cross validation to get the out-of-sample predictions. In the second layer, use ridge regression as stacking's combiner algorithm to combine the forecasting results of the first layer's models. The proposed model is evaluated on the load data collected from a transformer substation in Hangzhou, and is compared with the existing models based on Gradient Boosting Decision Tree, Long Short-Term Memory and CatBoost. The results demonstrate that the proposed model outperforms all the mentioned existing models.